jsilter
commented
3 months ago - Change setup.cfg so that we install torch+cuda based on platform_machine (architecture) not OS.
- Create sybil-predict CLI
- Change name from inference -> predict (seems to be more common). Rearrange imports so we don't need to install lifelines when just doing prediction.